From the course: NLP with Python for Machine Learning Essential Training
Unlock the full course today
Join today to access over 22,600 courses taught by industry experts or purchase this course individually.
Evaluate gradient-boosting model performance - Python Tutorial
From the course: NLP with Python for Machine Learning Essential Training
Evaluate gradient-boosting model performance
- [Instructor] We're starting to notice that model building is an iterative process. Unfortunately, we don't just write a couple lines of code, fit a model, and then wipe our hands of it. We start with an initial exploration, then we take what we learned, and we dive a little bit deeper to learn a little bit more. Then we take what we learned in that phase, and dive even deeper on a more specific model. Until eventually, we've a built a model that we're confident in, and that we thoroughly evaluated. With that context, we're going to enter the next phase of our build using GridSearchCV to explore our model, while allowing our prior exploration to point us in the right direction. So just to recap very quickly. GridSearch means setting up different parameter settings that you want to test, and then exhaustively searching that entire grid to determine the best model. And cross-validation means you take your dataset, divide it into k subsets, and then you repeat the holdout test method…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Contents
-
-
-
-
-
-
-
(Locked)
What is machine learning?4m 2s
-
(Locked)
Cross-validation and evaluation metrics7m 48s
-
(Locked)
Introducing random forest3m 4s
-
(Locked)
Building a random forest model8m 11s
-
(Locked)
Random forest with holdout test set12m 2s
-
(Locked)
Random forest model with grid search8m 48s
-
(Locked)
Evaluate random forest model performance8m 44s
-
(Locked)
Introducing gradient boosting4m 13s
-
(Locked)
Gradient-boosting grid search9m 44s
-
(Locked)
Evaluate gradient-boosting model performance9m 32s
-
(Locked)
Model selection: Data prep8m 25s
-
(Locked)
Model selection: Results9m 52s
-
(Locked)
-